Macintosh SE/30 logicboard recreation (thread revival)

Another thing is, it would be good to know if you’re getting any sad chimes, as would tell you if the ROM is being accessed, but without the speaker jack soldered in, audio won’t get routed to the internal speaker header. So either add that in, or solder bridge the two pads on the underside of where that connector goes (and make sure an internal speaker is connected).

IMG_7328.jpeg
 
Did you use the bitswapped images on Bolle’s GitHub? The Reloaded board needs byte-reversed code and this could potentially trip some people up.
Thanks for your replies! Very much appreciated!

Yes I directly burned the ROM images from the Github repo without any modification.
Then read them back to verify checking the check sum hashes which matched.

I did this for my first two boards and it’s a great idea for testing chips, although I’ve damaged a few cosmetically during removal even with a chip puller, so take care with that.
The chips need to be very clean before they’re inserted, so that’s one thing to check - one of them might not be making perfect contact.
Each chip was carefully removed and then painstakingly cleaned.
First I removed residue and buildup by very gently scraping pins which were effected. Then fluxed some braid and cleaned each until no access solder was apparent and they were very reasonably silver in color and uniformly smooth.
I then rinsed with ISO and brushed thoroughly with a toothbrush to clean.
Very carefully I pried each pin out a fraction from the underneath of the IC by very carefully inserting and levering a suitable tweezer tip in the gap between where the leg ends and the cavity that constrains their position. This ever so slightly shifted the pins outward and approximately back to their original position after being slightly compressed by the process of removal and cleaning.

Another thing is, it would be good to know if you’re getting any sad chimes, as would tell you if the ROM is being accessed, but without the speaker jack soldered in, audio won’t get routed to the internal speaker header. So either add that in, or solder bridge the two pads on the underside of where that connector goes (and make sure an internal speaker is connected).

View attachment 92472

Thanks for the advice and good catch. But this is a slightly older photo of my board and have since soldered this and tested with a speaker and I get no bong what-so-ever. Which means that the boot is not even getting to ROM or not even far enough into testing.
I've checked the RESET line and it goes HIGH shortly after starting and stays there.
I've check the oscillators and am getting ~32Mhz on the main and ~16Mhz on the CPU and ~32khz on the RTC.
I've checked voltages and they are very good just over 5V and pretty much right on 12V. And -12V is also present and strong.
As for the VRAM, if you look closely it's almost correct. I would have thought uninitialised RAM would have been far more random than this pattern - it's like it trying to draw the checker pattern but something isn't fetching from the correct spot, like bit-0 looks "out"
Anyway I had ordered some new Video RAM which should be arriving soonish.

Also here are the EPROMs I am using - compatible?
 

Attachments

  • tempImageez50yY.png
    tempImageez50yY.png
    2.7 MB · Views: 6
And as a bonus I have been working on removing components from a second original SE/30 board - the utility of the sockets now coming into play by being able to test another set of ICs (that is assuming the sockets are helping and not hindering).
I think if I had a known good set of ICs I would just solder them in directly for reliability and just keep sockets for the DIP ICs and CPU.
 
As for the VRAM, if you look closely it's almost correct. I would have thought uninitialised RAM would have been far more random than this pattern - it's like it trying to draw the checker pattern but something isn't fetching from the correct spot, like bit-0 looks "out"

Nope, this is the standard display appearance of uninitialized NEC 41264 chips. Other brands show stripes or other patterns.

Anyway I had ordered some new Video RAM which should be arriving soonish.

Also here are the EPROMs I am using - compatible?

I think so, they’re 5V 64K x 8bit chips, which sounds right to me.
 
Okay got some new VRAM, EPROM. Still getting not much other than horizontal stripes.
To program the EPROM I doubled each ROM from the GitHub repo:
cat u1 u1> u1_128k
cat u2 u2> u2_128k
etc
and burned and installed into sockets

IMG_3861.jpeg
View attachment IMG_3862.jpeg
 

Attachments

  • IMG_3858.jpeg
    IMG_3858.jpeg
    103.5 KB · Views: 1
I did some scoping and found that ROMN (ROM enable, pin 27) of the GLUE stays high. This goes to the 29F010’s output enable (pin 24) active low. The ROMs never output anything! Which is true; the address lines initially get some activity but not the data bus.
i also sounded out all the address lines from the GLUE to the 68030 and that was fine.

Anyone in Australia got a socketed SE/30 reloaded board? 😀
Would love to test my PLCC ICs.
 
I'm currently having this same issue with ROMN staying high, and trying to figure out what criteria determine the level of that pin (if GLUE does some kind of logic or just forwards it from somewhere else) or where to walk back on the schematic to check my work. I am not using PLCC sockets or flash, but a slot with a known good, original ROM. I get a default VRAM pattern and stable sync like you have, too.

I haven't yet found anything amiss with the reset circuit, though I've heard it's worth double-checking each test point's frequency for abnormalities, and checking that it goes low everywhere when the button is down. Every reset pin on mine appears to be the 5V it should, when the board is idle. If there's a dead component at play here, I pray it's not GLUE.
 
I did some scoping and found that ROMN (ROM enable, pin 27) of the GLUE stays high. This goes to the 29F010’s output enable (pin 24) active low. The ROMs never output anything! Which is true; the address lines initially get some activity but not the data bus.
i also sounded out all the address lines from the GLUE to the 68030 and that was fine.

Anyone in Australia got a socketed SE/30 reloaded board? 😀
Would love to test my PLCC ICs.

At this point would it not be easier to just obtain a ROM socket and use a known working SIMM?
 
I'd agree with this, unless you have a personal preference or special use case for the PLCC sockets, or another machine/SIMM adapter to test them in and rule them out. I'd also tried building mine with the sockets, but realized that since there's a proper coin cell clip for a less volatile battery, one would not need a hacked ROM anyway, as PRAM/XPRAM persisting correctly would allow 32-bit addressing to stay enabled across reboots, so Force32 wouldn't need to reset the machine every time.

That's just for modified ROMs and 32-bit cleanliness, though. Regardless of ROM choice, I think having one of those SIMM slot adapters with the sockets on the stick itself seems like a better idea if you want to use PLCCs. It adds complexity, sure, but that way you don't have to carefully press down pull up against your entire logic board if you ever want to take chips out. Releasing a SIMM from the slot takes no excess force or aim, and isn't liable to accidentally mess up any pins. The breakout/module would be much easier to remove PLCCs from once it was removed from the slot, and if anything did go wrong, it'd be a much easier repair or replacement.

To reiterate, though, that's only my personal rationale. If you prefer the sockets right on the board and want something different, go for it.
 
Back
Top